DETAILED ACTION
In response to communication filed on 01 December 2025, claims 1-12 are canceled. Claims 13-32 are newly added claims. Claims 13-32 are pending.
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Arguments
Applicant’s arguments, see “Claim Interpretation” filed 01 December 2025, have been carefully considered. Based on the canceled claims, the claim interpretation does not apply anymore.
Applicant’s arguments, see “Claim Objections” filed 01 December 2025, have been carefully considered. Based on the canceled claims, the claim objections haven been updated below.
Applicant’s arguments, see “Rejections under 35 U.S.C. § 112(a)” filed 01 December 2025, have been carefully considered. Based on the canceled claims, the claim rejections are withdrawn.
Applicant’s arguments, see “Rejections under 35 U.S.C. § 112(b)” filed 01 December 2025, have been carefully considered. Based on the canceled claims, the claim rejections are withdrawn.
Applicant’s arguments, see “Rejections under 35 U.S.C. § 101” filed 01 December 2025, have been carefully considered. Based on the canceled claims, the claim rejections are withdrawn.
Applicant’s arguments, see “Rejections under 35 U.S.C. § 103” filed 01 December 2025, have been carefully considered. The arguments are related to newly added claim limitations and are addressed in the rejection below.
Claim Objections
Claims 14, 18, 22, 28 and 31 are objected to because of the following informalities:
Claims 14 and 31 recite “retrieving, via a database query, segment metadata” should read as -- retrieving, via a database query, the segment metadata -- as it appears to be a typographical error and may cause antecedent basis issue.
Claim 18 recites “comprising segment metadata” should read as -- comprising the segment metadata -- as it appears to be a typographical error and may cause antecedent basis issue.
Claims 22 and 28 recite “comprises segment metadata” should read as -- comprises the segment metadata -- as it appears to be a typographical error and may cause antecedent basis issue.
Appropriate corrections are required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 13-18, 21, 24, 27 and 30-32 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnan et al. (US 12,008,309 B1, hereinafter “Krishnan”) in view of Stollman (US 2010/0074524 A1, hereinafter “Stollman”) further in view of Tran (US 2023/0252224 A1, hereinafter “Tran”).
Regarding claim 13, Krishnan teaches
A computer-implemented method (see Krishnan, [col 18 lines 8-10] “The methods can also be embodied as computer-usable instructions stored on computer storage media”) for generating a machine-generated document having a given document type, the method comprising: (see Krishnan, [col 20 lines 20-27] “for generating a unique copy is provided… set of documents is accessed. The set of documents may have a same author… The set of documents may contain documents of a single file type, or may comprise documents of various file types”).
receiving, an original document (see Krishnan, [col 20 lines 20-27] “for generating a unique copy is provided… set of documents is accessed. The set of documents may have a same author… The set of documents may contain documents of a single file type, or may comprise documents of various file types”; [col 6 lines 35-48] “an original document, such as document 200, may be any document type that conveys content therein, such as text, images, tables, graphs, and so forth… original documents can include various file types, such as JPEG (joint photographic experts group), GIF (Graphics interchange format), SVG (scalable vector graphics), PNG (portable network graphic), BMP (bitmap), TIFF (tagged image file format), PDF (portable document format), Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT, WPD), PowerPoint (e.g., PPT, PPTX), ODP (open document presentation), KEY (Keynote file), message file (MSG), email (EML), and other like document types”) by one or more processors of a computing system and (see Krishnan, [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”) via a communication interface,… (see Krishnan, [col 21 lines 59-60] “where tasks are performed by remote-processing devices that are linked through a communications network”) specifying the document type; (see Krishnan, [col 20 lines 20-27] “for generating a unique copy is provided… set of documents is accessed. The set of documents may have a same author… The set of documents may contain documents of a single file type, or may comprise documents of various file types”; [col 6 lines 35-48] “an original document, such as document 200, may be any document type that conveys content therein, such as text, images, tables, graphs, and so forth… original documents can include various file types, such as JPEG (joint photographic experts group), GIF (Graphics interchange format), SVG (scalable vector graphics), PNG (portable network graphic), BMP (bitmap), TIFF (tagged image file format), PDF (portable document format), Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT, WPD), PowerPoint (e.g., PPT, PPTX), ODP (open document presentation), KEY (Keynote file), message file (MSG), email (EML), and other like document types”).
generating, by the one or more processors, text content for a plurality of segments based on the document type, wherein each segment is generated by the one or more processors through automated retrieval and (see Krishnan, [col 6 lines 35-57] “original documents can include various file types, such as… Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT,WPD)… and other like document types… illustrates example terms within document 200… terms can include one or more words, such as a single word or phrase… term 202 is an introductory phrase, while term 204 is a single word. A term may also include one or more full sentences, such as term 206. While document 200 is illustrated as an mail, it will be appreciated that this is just one example of an original document, and that the technology will apply to other document types as well”; [col 7 lines 55-65] “term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document. Generally, NLP models may employ optical character recognition to identify characters… BERT comprises a text prediction component that aids in determining term phrases”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory” – Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments).
extracting,… executed by the one or more processors, defined terms from the generated text content,… (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”) to identify terms and (see Krishnan, [col 7 lines 47-52] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document. The terms extracted by term extractor 112 are candidate terms for modifying to generate unique copies… term extractor 112 extracts terms subject to term extraction rules 134”) definitions of the terms across the plurality of segments; (see Krishnan, [col 9 lines 57-66] “Having identified and extracted terms from a document, encoder 110 may employ alternative term determiner 114 to determine alternative terms for the extracted terms. In doing so, alternative term determiner 114 may use alternative term model 122. In general, alternative term model 122 receives an input term, and from the input term, outputs alternative terms. The output terms may have the same or semantically similar meaning as the input term… alternative term model 122 can be a neural network trained to identify alternative terms”; [col 11 lines 6-13] “Input extracted term 502 comprises "As many of you are aware," and responsive to this, the output set of alternative terms 508 comprises "As you may be aware," "As some of you are aware," "As many of you may know," "As was previously explained to you," "As you've been told," "As someone told you," and "As you might have heard." Each of the alternative terms has a same or similar meaning as extracted term 502”- Fig. 5 – the different areas on document 200 that identifies terms 502, 504, etc. have been interpreted as segments).
modifying, by a consistency checker module executed by the one or more processors, the defined terms to create consistent versions through computerized comparison of the definitions across the plurality of segments (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term 506”; [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”) and algorithmic generation of unified definitions (see Krishnan, [col 13 lines 55-62] “alternative terms selected at the computing device can be paired with the original extracted term as a term-alternative term pair… Term-alternative term pairs can be labeled and stored as tone model training data 138”) performed by the consistency checker module (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules) based on stored digital data structures in memory of the computing system; (see Krishnan, [col 13 lines 55-62] “alternative terms selected at the computing device can be paired with the original extracted term as a term-alternative term pair… Term-alternative term pairs can be labeled and stored as tone model training data 138”).
replacing, by the one or more processors, the defined terms in the text content with the consistent versions within a digital document data structure stored in the memory of the computing system to generate the machine-generated document; and (see Krishnan, [col 4 lines 49-51] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies”; [col 5 lines 4-7] “To generate a unique copy, an alternative term from the one or more alternative terms is selected and replaces the extracted term from the original document. This can be done for any number of terms throughout the original document”; [col 14 lines 17-25] “the term "As many of you are aware" in document 200 has been replaced with "You may be aware" in unique copy 802. The term '"hope' to finalize" has been replaced with "anticipate finalizing," while the term "I appreciate all of your hard work and efforts" has been replaced with "All of your hard work and efforts are appreciated." In doing so, unique copy generator 118 generates a unique copy of document 200”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”).
outputting, by the one or more processors, the machine-generated document to a user interface or external device via the communication interface (see Krishnan, [col 7 lines 10-18] “Unique copies can be distributed to individual recipients. Thus, each recipient receives a unique copy of the original document that is unique to the recipient. Unique copies may be provided in any manner, such as a printed document, an email attachment, a message body, or other like delivery method. A mapping (e.g., a data index) can be kept to indicate an association between a unique copy and a recipient, thus allowing identification of a recipient via the mapping when the unique copy is known”).
Krishnan does not explicitly teach receiving, a document request; processing of segment metadata retrieved from a document library of previously analyzed documents stored in a database; extracting, by a large language model; the large language model being configured.
However, Stollman discloses document analysis and teaches
receive a document request (see Stollman, [0033] “The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”).
processing of segment metadata retrieved from a document library of previously analyzed documents stored in a database; (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of receiving document request, retrieving segment metadata, processing segment metadata, document metadata, linking, prompt, template, before events, analyze terms, iterating through documents and after events as being disclosed and taught by Stollman, in the system taught by Krishnan to yield the predictable results of effectively analyze documents to perform relevant actions (see Stollman, [0041] “Any of the actions and/or options (which can be actions) performed (and/or triggered) by the document analysis engine 110 with respect to the document 164 can be applied to the target content as well… the document analysis engine 110 (e.g., the template module 160 of the document analysis engine 110) can, for example, block and/or allow access (by the requesting entity 140 or a different entity) to one or more portions of the target content based on processing of the document 164”).
The proposed combination of Krishan and Stollman does not explicitly teach extracting, by a large language model; the large language model being configured.
However, Tran discloses generate document by prompt-engineering and teaches
generate text by a large language model (see Tran, [0456] “The large language models can be used to generate text”).
the large language model being configured to generate text (see Tran, [0456] “The large language models can be used to generate text”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of large language model, database query, generate document outline data structure, content identifier, map data structure, comparing data and inconsistencies as being disclosed and taught by Tran, in the system taught by the proposed combination of Krishnan and Stollman to yield the predictable results of significantly speedup document generation with reduced cost (see Tran, [0095] “Advantages of the system may include one or more of the following. The system reduces the cost of writing documents by serving as writing assistants that fill (or in-between) details based on the abstract… The result is a significant speedup in document generation, while cost is reduced”).
Claims 24 and 30 incorporate substantively all the limitations of claim 13 in a computer-readable medium form (see Krishnan, [col 22 lines 16-28] “Computing device 1200 typically includes a variety of computer-readable media… Computer storage media, also referred to as a communication component, includes both volatile and nonvolatile, removable and non-removable media implemented… as computer-readable instructions, data structures, program modules, or other data”) and system form (see Krishnan, [col 21 lines 61-64] “computing device 1200 includes bus 1202, which directly or indirectly couples the following devices: memory 1204, one or more processors 1206, one or more presentation components 1208”; [col 23 lines 59-65] “various functions may be carried out by a processor executing computer-executable instructions stored in memory, such as database 106. Moreover, functions of encoder 110, alternative term model trainer 128, tone model trainer 130, and decoder 132, among other functions, may be performed by server 102, computing device 104”) and are rejected under the same rationale.
Regarding claim 14, the proposed combination of Krishnan, Stollman and Tran teaches
wherein generating text content for the plurality of segments comprises: (see Krishnan, [col 6 lines 35-57] “original documents can include various file types, such as… Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT,WPD)… and other like document types… illustrates example terms within document 200… terms can include one or more words, such as a single word or phrase… term 202 is an introductory phrase, while term 204 is a single word. A term may also include one or more full sentences, such as term 206. While document 200 is illustrated as an mail, it will be appreciated that this is just one example of an original document, and that the technology will apply to other document types as well”; [col 7 lines 55-65] “term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document. Generally, NLP models may employ optical character recognition to identify characters… BERT comprises a text prediction component that aids in determining term phrases”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory” – Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments).
retrieving, via a database query, (see Tran, [0307] “Users often query a search engine”) segment metadata from the document library based on similarity between the requested segment and previously analyzed segments; and (see Stollman, [0023] “The parsing module 150 can be configured to interpret (e.g., determine a meaning for, extract a meaning from) the document 164 based on one or more document parsing functions… the document parsing function 152 can be configured to parse at least a portion of text and/or a portion of a media object (e.g., an image, a video, an audio snippet) from the document 164. The document parsing function 152 can be configured to interpret one or more portions, or groups of one or more portions, parsed from the document 164”; [0026] “the document parsing function 152 can be configured to prepare the document for comparison with the template 162… portions of the document 164 can be rearranged by the parsing module 150 based on semantic relationships between words and/or the relationships between sections within the template 162… the template 162 and the document parsing function 152 can be defined so that they are compatible with (e.g., consistent with, customized for) one another”).
generating the text content (see Krishnan, [col 6 lines 35-57] “original documents can include various file types, such as… Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT,WPD)… and other like document types… illustrates example terms within document 200… terms can include one or more words, such as a single word or phrase… term 202 is an introductory phrase, while term 204 is a single word. A term may also include one or more full sentences, such as term 206. While document 200 is illustrated as an mail, it will be appreciated that this is just one example of an original document, and that the technology will apply to other document types as well”; [col 7 lines 55-65] “term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document. Generally, NLP models may employ optical character recognition to identify characters… BERT comprises a text prediction component that aids in determining term phrases”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory” – Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments) based on the retrieved segment metadata (see Stollman, [0023] “The parsing module 150 can be configured to interpret (e.g., determine a meaning for, extract a meaning from) the document 164 based on one or more document parsing functions… the document parsing function 152 can be configured to parse at least a portion of text and/or a portion of a media object (e.g., an image, a video, an audio snippet) from the document 164. The document parsing function 152 can be configured to interpret one or more portions, or groups of one or more portions, parsed from the document 164”; [0026] “the document parsing function 152 can be configured to prepare the document for comparison with the template 162… portions of the document 164 can be rearranged by the parsing module 150 based on semantic relationships between words and/or the relationships between sections within the template 162… the template 162 and the document parsing function 152 can be defined so that they are compatible with (e.g., consistent with, customized for) one another”). The motivation for the proposed combination is maintained.
Claim 31 incorporates substantively all the limitations of claim 14 in a system form and is rejected under the same rationale.
Regarding claim 15, the proposed combination of Krishnan, Stollman and Tran teaches
further comprising: generating, (see Tran, [0055] “to generate a document using computer-generated outlines, or alternatively using digitized hand-crafted outlines or storyboards”) by a document outlining module (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules) executed by the one or more processors, (see Krishnan, [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”) a document outline data structure (see Tran, [0055] “to generate a document using computer-generated outlines, or alternatively using digitized hand-crafted outlines or storyboards”) specifying the plurality of segments (see Krishnan, [col 19 lines 13-16] “The unique copy may be generated as part of a plurality of unique copies that each comprise distinct terms based on replacing alternative terms”) before (see Stollman, [0071] “the contract parsing function can be selected based on the contract type (shown at 530) before the contract template is selected”) generating the text content for the plurality of segments (see Krishnan, [col 6 lines 35-57] “original documents can include various file types, such as… Word document (e.g., DOC, DOCX), HTML (hypertext markup language), spreadsheets (e.g., XLS or XLSX), text files (e.g., TXT,WPD)… and other like document types… illustrates example terms within document 200… terms can include one or more words, such as a single word or phrase… term 202 is an introductory phrase, while term 204 is a single word. A term may also include one or more full sentences, such as term 206. While document 200 is illustrated as an mail, it will be appreciated that this is just one example of an original document, and that the technology will apply to other document types as well”; [col 7 lines 55-65] “term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document. Generally, NLP models may employ optical character recognition to identify characters… BERT comprises a text prediction component that aids in determining term phrases”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory” – Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments). The motivation for the proposed combination is maintained.
Claim 32 incorporates substantively all the limitations of claim 15 in a system form and is rejected under the same rationale.
Regarding claim 16, the proposed combination of Krishnan, Stollman and Tran teaches
wherein the consistency checker module is configured to: (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules).
analyze terms associated with the document (see Stollman, [0035] “based on analysis of an identifier(s) and/or a term(s) associated with the document 164”) the defined terms from the plurality of segments; (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction” - Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments).
identify inconsistencies between definitions of the defined terms across the plurality of segments; and (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term” - Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments).
generate the consistent versions by addressing the inconsistencies (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term”). The motivation for the proposed combination is maintained.
Regarding claim 17, the proposed combination of Krishnan, Stollman and Tran teaches
further comprising: generating a dedicated defined terms segment comprising the consistent versions and associated definitions; and (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term 506”; [col 4 lines 49-51] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies”).
inserting the dedicated defined terms segment into the machine-generated document (see Krishnan, [col 4 line 49 – col 5 line 2] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies… thus identifying which of the alternative terms have a common tone with the term from the original document… a selection is made of one or more alternative terms from the set of alternative terms based on tone”).
Regarding claim 18, the proposed combination of Krishnan, Stollman and Tran teaches
wherein the document library comprises a plurality of previously analyzed documents and, for at least one of the plurality of previously analyzed documents, structured metadata comprising (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function 152… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”) segment metadata and (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164”) document metadata (see Stollman, [0031] “accompanied by information defined based on the processing of the document 164 by the document analysis”). The motivation for the proposed combination is maintained.
Regarding claim 21, the proposed combination of Krishnan, Stollman and Tran teaches
wherein replacing the defined terms comprises: (see Krishnan, [col 4 lines 49-51] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies”; [col 5 lines 4-7] “To generate a unique copy, an alternative term from the one or more alternative terms is selected and replaces the extracted term from the original document. This can be done for any number of terms throughout the original document”; [col 14 lines 17-25] “the term "As many of you are aware" in document 200 has been replaced with "You may be aware" in unique copy 802. The term '"hope' to finalize" has been replaced with "anticipate finalizing," while the term "I appreciate all of your hard work and efforts" has been replaced with "All of your hard work and efforts are appreciated." In doing so, unique copy generator 118 generates a unique copy of document 200”; [col 18 lines 6-8] “various functions can be carried out by a processor executing instructions stored in memory”).
generating a mapping data structure that maps word (see Tran, [0099] “it will map the word into the embedding space based on its context”) each extracted defined term (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) to embedding space based on its context (see Tran, [0099] “it will map the word into the embedding space based on its context”) a corresponding consistent version and (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term 506”) an updated definition; (see Krishnan, [col 14 lines 41-42] “the number of changes that are made by modifying terms”).
comparing the first data (see Tran, [0254] “comparing the agent data and the caller data”) at least one defined term (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) comparing the first data with second data (see Tran, [0254] “comparing the agent data and the caller data”) a corresponding consistent version (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term 506”) to identify changes made during consistency checking; (see Krishnan, [col 44-46] “These changes can be used to identify the source of a document by identifying the individual changes made to the documents”).
replacing definitions associated with the at least one defined term (see Krishnan, [col 19 lines 14-15] “each comprise distinct terms based on replacing alternative terms”) when inconsistencies exist between (see Tran, [0469] “includes a thorough analysis of the language used, including any ambiguous or unclear terms, repetitive phrases, and inconsistent language”) an initial definition and the updated definition; and (see Krishnan, [col 14 lines 43-47] “there are three sets of alternative terms respectively for the three extracted terms. In this example, there are five options for extracted term 502 when generating a unique copy, those in alternative terms 610 and the original extracted term”).
iterating (see Stollman, [0036] “the iterative processing by the document parsing function 152 can be performed based on different document parsing functions (not shown) during different iterations”) through each segment (see Krishnan, Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments) to replace at least one occurrence of the at least one defined term with the corresponding consistent version (see Krishnan, [col 4 lines 49-51] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies”; [col 5 lines 4-7] “To generate a unique copy, an alternative term from the one or more alternative terms is selected and replaces the extracted term from the original document. This can be done for any number of terms throughout the original document”; [col 14 lines 17-25] “the term "As many of you are aware" in document 200 has been replaced with "You may be aware" in unique copy 802. The term '"hope' to finalize" has been replaced with "anticipate finalizing," while the term "I appreciate all of your hard work and efforts" has been replaced with "All of your hard work and efforts are appreciated." In doing so, unique copy generator 118 generates a unique copy of document 200”) in the mapping data structure (see Tran, [0099] “it will map the word into the embedding space based on its context”). The motivation for the proposed combination is maintained.
Claim 27 incorporates substantively all the limitations of claim 21 in a computer-readable form and is rejected under the same rationale.
Claims 19 and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnan, Stollman and Tran in view of Marian et al. (US 9,298,696 B2, hereinafter “Marian”).
Regarding claim 19, the proposed combination of Krishnan, Stollman and Tran teaches
wherein extracting the defined terms comprises: (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”).
… a defined term, (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) a corresponding definition, and (see Krishnan, [col 9 lines 57-66] “Having identified and extracted terms from a document, encoder 110 may employ alternative term determiner 114 to determine alternative terms for the extracted terms. In doing so, alternative term determiner 114 may use alternative term model 122. In general, alternative term model 122 receives an input term, and from the input term, outputs alternative terms. The output terms may have the same or semantically similar meaning as the input term… alternative term model 122 can be a neural network trained to identify alternative terms”; [col 11 lines 6-13] “Input extracted term 502 comprises "As many of you are aware," and responsive to this, the output set of alternative terms 508 comprises "As you may be aware," "As some of you are aware," "As many of you may know," "As was previously explained to you," "As you've been told," "As someone told you," and "As you might have heard." Each of the alternative terms has a same or similar meaning as extracted term 502”) a content identifier (see Tran, [0285] “with their unique identifiers which may be used to help describe the content”) linking (see Stollman, [0033] “an identifier, such as an electronic tag ( e.g., metadata, a link)”) the defined term (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) to a source segment; and (see Krishnan, [col 2 lines 44-46] “can be used to identify the source of a document by identifying the individual changes made to the documents”).
for the defined terms in multiple segments,… (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction” - Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments) from a subsequent segment… (see Krishnan, Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments – 204 is a subsequent segment).
The proposed combination of Krishnan, Stollman and Tran does not explicitly teach generating a dictionary data structure comprising entries, wherein each entry comprises a defined term, a corresponding definition; updating an existing entry in the dictionary data structure by adding additional definition information from a subsequent segment when the additional definition information adds detail not present in the existing entry.
However, Marian discloses dictionary and teaches
generating a dictionary data structure comprising entries, wherein each entry comprises concepts associated with the components (see Marian, [col 4 lines 19-22] “The domain dictionaries may thus include entries for the concepts that are associated with the components used to model the domain and one or more definitions for those concepts”; [col 22 lines 24-29] “may determine whether to generate one or more mapping dictionaries for the concept. If so, the computing device may generate the mapping dictionary in step 1316. The mapping dictionaries may be used to map an application-specific concept name to a domain dictionary name”).
updating an existing entry in the dictionary data structure by adding additional definition information… when the additional definition information adds detail not present in the existing entry (see Marian, [col 21 lines 23-61] “The concept name may be, for example, “PartyID.” In step 1304, the computing device may determine whether the received concept name is an update to a concept existing, for example, in the domain dictionary. The computing device may make this determination based on the unique concept ID associated with the received concept name… If the concept name is not an update to an existing concept (step 1304: No), the computing device may determine that the concept is new and proceed to create a new entry for the domain dictionary. The computing device may make this determination if the user inputs a concept name that is not associated with a concept ID. In step 1308, the computing device may generate a concept ID for the new concept…the computing device may generate a concept definition for the new concept”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of dictionary data structure as being disclosed and taught by Marian, in the system taught by the proposed combination of Krishnan, Stollman and Tran to yield the predictable results of effectively mapping information in dictionary (see Marian, [col 2 lines 39-48] “an application uses a first concept name from an application dictionary to describe an event in response to a determination that the event occurs at the application. An entry for the event may be added to a data log for the application. The entry may include the first concept name from the application dictionary. The method may additionally comprise generating a mapping of the first concept name from the application dictionary to a second concept name from a domain dictionary. The domain dictionary may be different from the application dictionary”).
Claim 25 incorporates substantively all the limitations of claim 19 in a computer-readable form and is rejected under the same rationale.
Claims 20 and 26 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnan, Stollman and Tran in view of Rahmani et al. (US 11,934,801 B2, hereinafter “Rahmani”).
Regarding claim 20, the proposed combination of Krishnan, Stollman and Tran teaches
wherein modifying the defined terms comprises: (see Krishnan, [col 14 lines 41-42] “the number of changes that are made by modifying terms”).
inserting the defined terms into (see Krishnan, [col 4 line 49 – col 5 line 2] “The set of alternative terms includes candidate terms that can be used to replace the term in the original document when generating unique copies… thus identifying which of the alternative terms have a common tone with the term from the original document… a selection is made of one or more alternative terms from the set of alternative terms based on tone”) a pre-defined prompting template to generate a defined terms prompt; (see Stollman, [0034] “the document parsing functions 152 (or a portion thereof) used by the parsing module 150 and/or the template 162 (or a portion thereof) used by the template module 160 to process the document 164 can be selected based on a document type associated with the document 164 based on analysis of an identifier and/or a term from the document 164… the template module 160 can be configured to apply one or more templates (such as template 162) based on the portion(s) of the document 164 being defined based on a EULA document type… the document analysis engine 110 can be configured to present an option (e.g., an option in a prompt) to the requesting entity 140 (or a different entity) to accept (e.g., confirm) or reject the selection (or portions of the selection) of the template(s) and/or document parsing function(s)”).
providing the defined terms prompt… (see Stollman, [0034] “the document parsing functions 152 (or a portion thereof) used by the parsing module 150 and/or the template 162 (or a portion thereof) used by the template module 160 to process the document 164 can be selected based on a document type associated with the document 164 based on analysis of an identifier and/or a term from the document 164… the template module 160 can be configured to apply one or more templates (such as template 162) based on the portion(s) of the document 164 being defined based on a EULA document type… the document analysis engine 110 can be configured to present an option (e.g., an option in a prompt) to the requesting entity 140 (or a different entity) to accept (e.g., confirm) or reject the selection (or portions of the selection) of the template(s) and/or document parsing function(s)”) the consistency checker module; and (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules).
generating the consistent versions (see Krishnan, [col 10 line 49 – col 11 line 2] “each of the extracted terms are provided as inputs to alternative term model 514… Responsive to the inputs, alternative term model 514 outputs a set of alternative terms, e.g., one or more alternative terms that have a same or similar meaning as the input terms… Other types of alternative terms include changes to the sentence structure…. a term that is a sentence may be rewritten from active voice to passive voice, or vice versa. For example, an alternative term for "the child broke the window" could be rewritten to "the window was broken by the child"… these are only some examples of alternative terms that may be generated”; [col 11 lines 42-50] “alternative terms, such as those generated by alternative term determiner 114, can be further processed for factual and grammatical accuracy… terms not meeting factual accuracy or grammatical correctness thresholds can be removed as candidate alternative terms”; [col 11 lines 40-41] “each of the alternative terms has a same or similar meaning compared to extracted term 506”) by the consistency checker module (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules) with the defined terms prompt as pre-prompt tuning (see Stollman, [0034] “the document parsing functions 152 (or a portion thereof) used by the parsing module 150 and/or the template 162 (or a portion thereof) used by the template module 160 to process the document 164 can be selected based on a document type associated with the document 164 based on analysis of an identifier and/or a term from the document 164… the template module 160 can be configured to apply one or more templates (such as template 162) based on the portion(s) of the document 164 being defined based on a EULA document type… the document analysis engine 110 can be configured to present an option (e.g., an option in a prompt) to the requesting entity 140 (or a different entity) to accept (e.g., confirm) or reject the selection (or portions of the selection) of the template(s) and/or document parsing function(s)”).
The proposed combination of Krishnan, Stollman and Tran does not explicitly teach prompt as a few-shot learning example to the consistency checker module.
However, Rahmani discloses few shot learning and teaches
provide prompt as a few-shot learning example to GPT-3 (see Rahmani, [col 11 lines 7-10] “GPT-3 facilitates a wide variety of tasks through few-shot learning. “Few-shot learning” refers to the fact that the completion predicted by the model can be tuned by providing only a few completion examples in the prompt”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of few shot learning as being disclosed and taught by Rahmani, in the system taught by the proposed combination of Krishnan, Stollman and Tran to yield the predictable results of effectively applying few-shot learning to pre-trained language models (see Rahmani, [col 11 lines 1-13] “Some pre-trained language models (PTMs) include… The predicted text tries to maintain the flow of the text in the prompt. GPT-3 facilitates a wide variety of tasks through few-shot learning. "Few-shot learning" refers to the fact that the completion predicted by the model can be tuned by providing only a few completion examples in the prompt; for present purposes, this means at most ten examples, but other contexts may specify a different limit for the "few" in "few-shot learning"”).
Claim 26 incorporates substantively all the limitations of claim 20 in a computer-readable form and is rejected under the same rationale.
Claims 22 and 28 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnan, Stollman and Tran in view of Meltzer et al. (US 2018/0293907 B2, hereinafter “Meltzer”).
Regarding claim 22, the proposed combination of Krishnan, Stollman and Tran teaches
wherein: the segment metadata comprises segment metadata generated (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”) by a metadata creator module (see Krishnan, [col 22 lines 37-38] “Communication media typically embodies… program modules” – there are plurality of modules) that analyzes text of previously analyzed documents (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”) divided into logical segments, and (see Krishnan, [0135] “data that is logically grouped or clustered to provide context”).
the segment metadata is stored in a store… (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”) a corresponding previously analyzed document stored in a data store (see Stollman, [0033] “an identifier, such as an electronic tag (e.g., metadata, a link) and/or one or more terms (e.g., a title/heading, a paragraph) associated with one or more portions of the document 164 (e.g., the entire document 164), can be received, interpreted, and used by the document analysis engine 110 to select the document parsing function… The identifier(s) and/or term(s) can be received by the document analysis engine 110 in response to a request from the document analysis engine 110 and/or extracted from one or more portions of the document 164 (and/or data associated with the document 164)”; [0079]-[0082] “the document analysis engine 710 can be configured to track (e.g., collect, store) information related to document analysis requests by entity Y, and can use that tracked data (also can be referred to as historical data) … The document analysis engine 710 can be configured to store the tracked data in, for example, a local memory (not shown) and/or a remote database… the document can be determined based on historical data stored at the document analysis engine”).
The proposed combination of Krishnan, Stollman and Tran does not explicitly teach a first data library having a link to a separate document library.
However, Meltzer discloses linking with library teaches
a first data library having a link to… a separate document library (see Meltzer, [0331] “a portion of the engagement data and metadata (validation metadata) that is linked to the engagement container”; [0345] “maintaining access to engagement data for electronic documents in the library, engagement data linked to the engagement containers in electronic documents in the library, and engagement data comprising parameters of an engagement based on the segments of content in the linked engagement containers”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of linked library as being disclosed and taught by Meltzer, in the system taught by the proposed combination of Krishnan, Stollman and Tran to yield the predictable results of effectively managing electronic document (see Meltzer, [0007] “A technology is described for managing engagement and presentation of an electronic document such as an e-book in a way that makes it both efficient on small platform devices, and engaging in a way that can encourage use of the devices”).
Claim 28 incorporates substantively all the limitations of claim 22 in a computer-readable form and is rejected under the same rationale.
Claims 23 and 29 are rejected under 35 U.S.C. 103 as being unpatentable over Krishnan, Stollman and Tran in view of Selbie et al. (US 2012/0271844 A1, hereinafter “Selbie”).
Regarding claim 23, the proposed combination of Krishnan, Stollman and Tran teaches
further comprising: prior to (see Stollman, [0071] “before the contract template is selected”) extracting defined terms, (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) iterating through document (see Stollman, [0036] “the iterative processing by the document parsing function 152 can be performed based on different document parsing functions (not shown) during different iterations”) the plurality of segments in sequence, (see Krishnan, Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments) wherein extracting defined terms comprises, (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”) for a first segment,… (see Krishnan, Fig. 2 – the different areas on document 200 that identifies terms 202, 204, etc. have been interpreted as segments – 202 has been interpreted as first segment) and for each subsequent segment, (see Krishnan, Fig. 2 – the different areas on document 200 that identifies terms 202, 204, 206, etc. have been interpreted as segments – 204 and 206 have been interpreted as subsequent segment) extracting defined terms from the subsequent segment and… (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”; Fig. 2 – the different areas on document 200 that identifies terms 202, 204, 206, etc. have been interpreted as segments – 204 and 206 have been interpreted as subsequent segment) after (see Stollman, [0035] “After the document parsing function 162”) the defined terms have been extracted from the plurality of segments (see Krishnan, [col 7 lines 47-58] “To identify terms in a document, such as term 202, term 204, and term 206 of document 200, encoder 110 may employ term extractor 112. In general, term extractor 112 identifies and extracts terms from an original document… term extractor 112 uses NLP model 120 to identify terms within a document. NLP model 120 may be a machine trained model that identifies and extracts text from the document”; [col 3 lines 29-31] “Such models can be used to paraphrase, i.e., generate a text fragment similar but not identical to an original fragment, textual documents”; [col 3 line 64 – col 4 line 2] “constraints that proper nouns are preserved in the paraphrased fragment, numbers are preserved in either decimal or literal form, and the paraphrased fragment does not differ too significantly from the original (e.g., by number of words), among other constraints”; [col 4 lines 35-39] “To make modifications to the text, terms are extracted from the original document. The terms may be individual words, phrases, or sentences. In some cases, the terms are extracted after applying an initial set of rules that exclude certain terms from being chosen for extraction”; Fig. 2 – the different areas on document 200 that identifies terms 202, 204, 206, etc. have been interpreted as segments).
The proposed combination of Krishnan, Stollman and Tran does not explicitly teach creating an empty defined terms data structure, adding the extracted defined terms to the defined terms data structure, and wherein modifying the defined terms is performed.
However, Selbie discloses providing relevant information and teaches
creating an empty defined terms data structure, (see Selbie, [0032] “an empty data structure can be created for storing respective sets of temporally recognized term”).
adding the extracted defined terms to the defined terms data structure, and wherein modifying the defined terms is performed (see Selbie, [0059] “one or more updated data structures can be respectively populated with an updated set of temporally recognized terms. For example, a new, empty bit array may be created, and populated with the updated terms comprised in the updated set”).
It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to include the functionality of creating empty data structure, modifying and adding information in the data structure as being disclosed and taught by Selbie, in the system taught by the proposed combination of Krishnan, Stollman and Tran to yield the predictable results of effectively managing electronic document (see Selbie, [0019] “that provides for quickly and efficiently finding keywords in a user message, such as one comprising an instant message, and identifying the keywords so that a user may find relevant information for a selected keyword. In this way, for example, a richer user experience can be provided for a user message, such as by allowing the user to select relevant information and add it to the user message ( e.g., embedding images, video, and reference information). Further, as an example, the user may be able to identify relevant information associated with a keyword that provides more detail about a particular topic, such as entertainment venues and times, which may be used to provide an improved user experience”).
Claim 29 incorporates substantively all the limitations of claim 23 in a computer-readable form and is rejected under the same rationale.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to VAISHALI SHAH whose telephone number is (571)272-8532. The examiner can normally be reached Monday - Friday (7:30 AM to 4:00 PM).
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, AJAY BHATIA can be reached at (571)272-3906. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/VAISHALI SHAH/Primary Examiner, Art Unit 2156